Rain shadow

Why can one side of a mountain range be rainforest and the other desert, under the same weather?

On the Olympic Peninsula the Hoh valley takes more than three metres of rain a year and grows temperate rainforest. Seventy-odd kilometres away, on the far side of the same mountains, Sequim takes something closer to four hundred millimetres and irrigates. Same latitude, same ocean, same weather systems arriving on the same day.

The easy answer — that the mountains "block" the rain — explains the dryness but not the other half of what is observed, which is that the lee side is also hotter. A wall casts a shadow; it does not warm what stands behind it. So what is the air actually doing as it crosses?

Follow one parcel of air in from the ocean. It meets the range and has nowhere to go but up. Rising, it moves into lower pressure and expands, and expanding costs it energy — so it cools. That is the ordinary business of the atmosphere, worked through under cooling with altitude; the rate for unsaturated air is close to 9.8 degrees per kilometre, fixed by thermodynamics rather than by weather.

Cool a parcel far enough and it reaches saturation, and here the arithmetic changes. Vapour condensing into droplets gives back the latent heat it absorbed when it evaporated off the sea. That release partly offsets the cooling of expansion, so from the condensation level upward the parcel cools more slowly — roughly 5 degrees per kilometre in warm, moist conditions, though the figure varies with temperature and approaches the dry rate in very cold air. Cloud forms; droplets grow; rain falls out and stays on the windward slope.

Now the crest, and the descent. The parcel is compressed as it sinks into higher pressure, and compression warms it. At what rate? Here is the hinge of the whole thing: its water fell out as rain on the way up, so there is nothing left to evaporate and no cooling to offset the compression. It warms at the full dry rate, all the way down.

So the parcel climbs at one rate and descends at a faster one. Put numbers on a crossing. Start at 20 degrees at the coast; the parcel cools at 9.8 per kilometre until it saturates at, say, a kilometre up, reaching about 10 degrees. From there to a 3-kilometre summit it cools at 5 per kilometre, arriving near freezing, shedding rain the whole way. Then it descends three kilometres at 9.8 per kilometre and reaches the lee floor at nearly 30 degrees — ten degrees warmer than it set out.

What has that done to the dryness? Two things at once, and this is why the contrast is so severe. The parcel carries far less water than it did, because most of it is now on the other side of the mountain. And warmer air can hold much more vapour than cool air, so the same shortfall counts for more: relative humidity collapses, and instead of raining, the descending air actively evaporates moisture out of soil and vegetation. The lee is not merely unwatered; it is being dried.

The pattern repeats wherever the geometry does: the Andes make the Atacama, the Sierra Nevada makes Death Valley, the Himalaya makes the Tibetan Plateau. When the descent is fast the warming has its own names — the föhn in the Alps, the chinook on the eastern slope of the Rockies, which can lift winter temperatures tens of degrees in hours.

A few refinements tie the steps together.

The asymmetry is not between "wet air" and "dry air" as such — it is between two lapse rates, and it only bites if condensation actually happens on the way up. Air that crosses a low ridge without ever saturating comes down at the temperature it went up with, and casts no rain shadow at all. Height matters because it sets whether the saturation threshold is crossed and how much of the climb happens above it.

It is also worth separating this from the other great dryness mechanism. The subtropical deserts examined under desert latitudes are made by air sinking on a planetary scale at the edge of the Hadley circulation, with no mountain involved. A rain shadow is the same physics of subsidence applied locally and forced by terrain, which is why it can make a desert at any latitude — and why some of the driest places on Earth, the Atacama among them, are where both mechanisms act on the same air.

One honest simplification: real air does not cross a range as a single sealed parcel. It mixes with its surroundings, some goes around rather than over, blocked flow can dam up on the windward side, and some condensate is carried across the crest and re-evaporated. The parcel calculation gives the direction and roughly the magnitude, not a forecast.

• Why the saturated lapse rate varies with temperature, and what that means for rain shadows in cold climates.
• How blocked and diverted flow around a range, rather than over it, changes where the rain actually falls.

Every term the collection defines is gathered in the glossary.